Tuesday, June 28, 2011

Responsible CIOs show restraint!

The HP Discover event is now history but the arrival by chance of a photo from an earlier ITUG event together with a posting in HP’s web site has me rethinking where NonStop is headed …

Mark Faithful forwarded me a photo this week that was taken at the ITUG event in San Jose, late October, 2002. It is of a luncheon held in honor of that year’s guest speaker, Gene Kranz whose book “Failure is not an Option” included a detailed account of the 1970 effort that went into rescuing the Apollo 13 crew. The photo above captures Kranz and his wife with the ITUG Summit 2002 committee.

Shortly after this photo was taken, Kranz addressed a full house of community members and he reflected on the events that led to the successful recovery of the crew. I have always enjoyed attending user events and have found participation highly rewarding. No matter where they may be held or how big or small an audience they attract, they have proven to be the best way to pick up on the latest trends concerning NonStop.

This post is a kind of milestone of sorts. Depending on who is doing the counting, this is my 200th post, and shortly I will begin my fifth year of blogging. No longer is this my sole presence in social networks as I now blog for a number of clients as well, but this blog remains the one that allows me to cover topics focused on the NonStop community and, much like The Connection magazine, I try to cover user deployments as often as I can - how many readers recall how this blog traces its roots back to the column, Real Time View, that for many years was a feature of The Connection magazine?

ITUG is now a part of the much bigger user community, Connect, and HP has steered the annual user gathering towards a much richer, multi-platform experience, where all of its products are showcased. While some attendees were confused to find sessions and group meetings in conflict with general sessions and keynotes, HP Discover still managed to pull together several lively Special Interest Group (SIG) meetings.

I was reminded of this aspect of HP Discover while scrolling through discussions on LinkedIn groups where one such entry pointed me to the report filed by Kevin McDowell of HP’s Total Customer Experience (TIC) group. McDowell posted a summary to the Mission Critical Computing blog, hosted on HP’s web site and can be found at:


“This year’s HP Discover NonStop SIG meeting,” McDowell opened with, before observing how “most (of the) discussion was around the marketing of the NonStop brand and offerings. NonStop has made significant in-roads in becoming an open / non-proprietary system and it appears we need to get the word out more effectively.”

If this doesn’t grab your attention, or have you raise an eyebrow, a little later McDowell asks “do you have examples of situations where NonStop was overshadowed by other marketing efforts? Or perhaps situations where NonStop was misrepresented as a dead-end product?” Now McDowell does admit in the post that he has been involved with NonStop for 23 years, which is about the same length of time as was my own involvement (I joined Tandem Computers in March, 1988), but even so, I think he may be onto something.

Could NonStop benefit from a greater marketing profile? Is the NonStop community becoming frustrated at seeing very little promotional activity in support of NonStop? And what’s this about NonStop being misrepresented as a dead-end product? Before it looks like I’m being a little too negative on this subject, dwelling too long on perceived failures from within HP, after reading McDowell’s summary I have to admit I like it that someone has at least started to think about how to bring better focus on to the NonStop product line.

For me, NonStop will never be general purpose ,and of that I have no reservations. It’s a specialty system as much as IBM’s mainframe, or even HP’s own SuperDome. And there’s a place for specialty systems within company data centers for many years to come, I believe, and that shouldn’t come as a surprise to any of my readers.

As I look at general purpose systems today, they are all dumbing down to run on inexpensive x86 chips supported by Windows, with some Linux and even Unix thrown into the mix. As workloads have grown and their users have tried to keep up by scaling out, they have become increasingly harder to manage. As one HP executive reminded me recently, have you ever tried to upgrade the firmware on a thousand Windows servers? Not cool …

It is this sense of despair (and failure) from trying to patch together clusters of inexpensive systems that has led to these very same users declaring that there must be a better way and they are now rushing to cloud computing, just the latest variant, or the service bureau of time sharing model many of us grew up with in the late ‘60s and early ‘70s. Sorry, the bits and pieces may be new with cloud computing but the model itself is pretty dated. And there’s only so much make-up you can apply to cover the wrinkles and hide all the blemishes.

CIOs will take advantage of the cloud computing model as, and when, it suits their needs and there will be a sizable move to cloud computing as a result. However, with the introduction of clouds, the necessity to look at how many core applications will move across will generate restraint on the part of responsible CIOs. There will be companies who will simply chose to run their own infrastructure and within this set of companies, specialty systems will retain a home, and among them NonStop will prevail.

Drilling deeper into this demographic, there will be even those companies where the history of NonStop in OLTP may encourage the pursuit of NonStop systems being deployed as essentially smart front-ends to clouds, private as well as public, in a manner not too dissimilar to how our early ATM networks were created, and for those attending HP Discover, a demonstration of configurations not too far removed from this model proved a popular draw card for the NonStop booth. Don’t rule out, either, the inroads NonStop will make into the high-end Oracle marketplace as HP further distances itself from the product offerings of a vendor it has long supported.

No, NonStop will not be general purpose and if this is the case, what then of marketing? What then of reversing any perception of NonStop being a dead-end product? Should we expect to see the new management team at HP get behind NonStop with a big-budget marketing campaign?

I recently watched the 70’s movie, 24 hours of LeMans. Apart from the comments coming from my wife, Margo, about how dreadfully boring the film is – no dialogue for the opening 40 minutes - it’s still one of the best motion pictures that captures the spirit of motor sport. In the film, the character, Michael Delaney, played by Steve McQueen, observes “racing’s important to men who do it well. When you’re racing, it ... it’s life. Anything that happens before or after... is just waiting.”

And I couldn’t help but see the similarity between these observations and what McDowell is trying to whip-up support for, when it comes to the NonStop server. “NonStop customers have always been a passionate group,” was how McDowell began his report following HP Discover and we are often accused of not having a life outside of NonStop! There are those companies after all, running NonStop and with the mission-critical applications they run, their companies are very much alive. And competitive, and with time to innovate.

For those companies, yet to understand the value proposition of NonStop, and who move from one model to the next looking to capitalize on the next big thing then no amount of marketing will deter them from going down that path. With little comprehension of what they are missing out on, well naturally, all I can suggest is that they too … are just waiting!

No comments: